INDEX

Explanations

routinely performed actions

The neuron flags adverbs (and related forms) that denote routine or habitual occurrence (e.g. “routinely,” “being,” “employed”).

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

-2.05

TypedDataSet

-1.88

戬

-1.80

 керами

-1.79

-1.73

to

-1.71

 incite

-1.70

 ignite

-1.70

ducati

-1.70

 lenient

-1.69

POSITIVE LOGITS

注意事

2.25

豋

2.22

俢

2.20

 scrit

1.97

玙

1.95

:</

1.93

1.91

 corrispond

1.90

࿚

1.89

number

1.88

Activations Density 0.013%